Procurement teams often struggle to fully leverage the valuable supplier knowledge captured during RFx (Request for Information, Proposal, or Quotation) events, mainly due to the high volume of unstructured data and fragmented data storage, which limits access to past insights. This paper explores how generative AI can preserve, retrieve, and apply RFx knowledge to strengthen supplier discovery, streamline event execution, and support data-driven strategic sourcing decisions. To address this challenge, the project developed a custom-built generative AI chatbot named Raffa, designed to extract meaningful insights from a wide range of historical RFx documents, and evaluated its performance alongside the sponsor’s internal GenAI Platform. Raffa was built on a Retrieval-Augmented Generation (RAG) framework integrating prompt engineering, structured metadata, and a vector database to enable context-rich retrieval. Both solutions were assessed against five criteria: truthfulness, answer accuracy, contextual relevance, handling of complex queries, and response structure. Results showed that metadata and prompt engineering significantly improved Raffa’s contextual relevance and overall response quality. Raffa outperformed the GenAI Platform in three of the five evaluation areas, particularly excelling in delivering precise and context-aware answers, even for complex queries. Additionally, the project introduced an RFx Knowledge Intelligence Framework and Prompt Library to guide scalable generative AI adoption across the RFx lifecycle.